Research Of Blast Furnace Expert System Based On Multi-model Integrated

Smelting in blast furnace is a complicated process under high temperature including a lot of physical change, chemical change and transmission process. In its long-term developing process, people gradually realized that using expert system to control the blast furnace is the development direction the blast furnace controls. The expert system which realized the closed loop control of blast furnace manifests maximum value in the practice.This article introduced a blast furnace expert system which based on the multi-model and elaborated the principle and method of it. On this basis we present a complete structure of the blast furnace expert system and focus introduced the abnormal states judgment systems.The abnormal states judgment systems is the important component of the blast furnace expert system for only under the stability states we just could optimize and predictive control the blast furnace smelting process.The blast furnace expert system which established before still have objections such as its real-time and learning ability is poor. While the fuzzy neural networks which combined fuzzy expert system with neural network fully absorb the advantages of fuzzy logic theory and neural network. The fuzzy neural networks could not only dealing with the expert knowledge and experience,but also could enhance judgment ability by self-learning. It have the advantages that the traditional expert system does not have. Introduced the fuzzy neural networks into expert system is an important research direction of blast furnace expert system.This paper elaborated the theoretical foundation,system structures and software processes of the abnormal states judgment systems detail and at last simulation it to test the feasibility of the method.